Perhaps one of the most vexing aspects of congenital cardiac disease is the current inability to explain its origin. Environmental causes have been invoked, and until recently only scant evidence had pointed toward a genetic component. Recent experimental data, combined with advances in human genetics, have now provided a clearer understanding of how some malformations may occur, and certainly have illuminated general concepts that are certain to apply to congenital cardiac disease in general. One of the most important developments has been the paradigm shift from grouping lesions based on clinical presentation to understanding the anomalies based on their embryonic and genetic origins. Thus it is now clear how an inherited mutation can result in a family with one individual having an interatrial communication and another tetralogy of Fallot, while still being considered the same genetic defect. This chapter reviews the various etiologies, environmental and genetic, with a constant eye toward the embryology of the heart, with the hope that synthesizing the current knowledge will provide a useful insight into the fundamental basis of congenital cardiac malformations.
Epidemiology of Heart Disease: Genes Versus Environment
The study of the etiology of congenital cardiac disease initially focused on epidemiologic studies, which mainly incorporated the identification of factors that influence the incidence of the various lesions. This is in large part because familial inheritance is not obvious, and thus a tractable focus is environmental influence and assessment of heritability. These studies primarily led to the conclusion that there were multifactorial influences. Several difficulties are apparent with these studies. First, intrauterine mortality due to congenital cardiac disease is difficult to assess, and conversely, in addition to the nearly 1% of children with cardiac malformations, an additional 1% to 2% of the population harbor more subtle cardiac developmental anomalies that only become apparent later in life. Second, familial associations are rarely obvious. In retrospect, this should be evident from the observations that defined mutations in a single gene can cause seemingly unrelated lesions, compounded by forme fruste or low genetic penetration.
Among epidemiologic studies, the Baltimore-Washington Infant Study was a prospective surveillance for live-born cases from 1981 to 1989, with a case-control study to determine etiologic associations. Over 4000 cases were identified among close to 1 million live births. Some clues about inheritance were obtained, but despite suggestive information, the focus of genetic evaluation was on chromosomal anomalies and known hereditable syndromes and not on identification of specific mutations. Thus, although several teratogenic causes of heart defects have been documented, the underlying mechanisms have not been elucidated. Most recently, prenatal use of angiotensin-converting enzyme inhibitors has been identified as a strong risk factor for congenital defects that include cardiac lesions.
Other than obvious associations with chromosomal syndromes, such as atrioventricular septal defects in the setting of Down syndrome, genetic causes have been slow in their discovery and characterization. Some cardiac conditions have a clear familial component, and include Marfan syndrome, Williams syndrome, and Holt-Oram syndrome. Recognition of the syndrome produced by deletion of chromosome 22q11 has caused a paradigm shift in how clinicians now think about the genetic contribution to congenital cardiac malformations. The deletion syndrome is associated with a host of cardiac lesions, ranging in severity and mostly involving the ventricular outflow tracts. The deletion has been reported in up to three-fifths of those with interrupted aortic arch, one-third of those with common arterial trunk, one-sixth of those with tetralogy of Fallot, and one-tenth of patients with ventricular septal defect. Routine testing is now standard, although its influence on cardiac outcomes is as yet unclear. It is interesting to note that the identification of the most likely causative gene came from studies in the mouse.
In contrast to clearly defined syndromes, most congenital cardiac malformations rarely occur in families with a sufficient number of affected members to lend themselves to genetic linkage analyses. Also, when familial cases occur, they are often marked by heterogeneity of defect, and affected family members may manifest as cardiomyopathy or arrhythmia rather than congenital malformations. Decreased penetrance and variable expressivity also occurs and suggests that additional environmental and genetic factors may contribute to risk of malformation. Historically, therefore, the risk factors for reoccurrence of the lesions were based on epidemiologic studies such as those described above and were classified broadly into family inheritance and vague environmental considerations. More recently, genetic analyses have determined that even some common types of defects have a genetic component. For example, atrial septal defects and aortic valves with two leaflets have been shown to be inheritable. The subsequent isolation of the causative genetic mutations substantiated this notion.
Cardiac Development: Key Concepts
While early anatomic descriptions have provided significant insights into normal cardiac development, modern genetic experimentations with model organisms have been particularly useful in deciphering the anatomic and genetic contortions that the developing heart must undergo to become a formed and functional organ. It has become clear that the genetic pathways that operate in such diverse species as the fruit fly, zebrafish, and mouse are relevant to each other and provide important biologic insights that are relevant to human disease. In particular, the fruit fly and the zebrafish have permitted the discovery of previously unknown pathways due to their use in large-scale phenotypic-based screens for discovery of genes. Similarly, genetic manipulation in the laboratory mouse, whose cardiovascular system is nearly identical to that of humans, has allowed profound insight into the mechanisms underlying human disease.
Origins of the Heart
The vertebrate heart arises from paired pools of mesodermal precursors. Cardiac differentiation begins shortly after gastrulation (the early stage of embryonic development when germ layers are established) has begun, and the first clear markers of the differentiating heart are apparent near the end of gastrulation, this occurs during the eighth day of development in the mouse. The process demarcates a horseshoe-shaped group of cells called the cardiac crescent (see also Chapter 3 ). The cells of the cardiac crescent come together at the midline of the embryos, where their fusion and anterior growth leads to the formation of the linear heart tube. This beating structure breaks the symmetry of the embryo, and loops toward the right side as distinct chambers form during the ninth day in the mouse. Looping proceeds during the 10th day, with growth of distinct chambers, giving rise to a heart composed of left and right atriums and ventricles. Subsequent steps in cardiac morphogenesis refine the distinctions between each chamber, and separate the left and right sides by growth of septums.
The outflow tract, the right ventricle, and a large component of the atria arise not from cardiac crescent–derived myocardium, but from a population of cardiac cells that form more anteriorly—the so-called second heart field. As defined by genetic lineage analysis, the second field originates from mesoderm expressing the transcription factor Isl1 near the area of the heart-forming mesoderm. More recently, clonal lineage tracing has revealed that the second heart field derivatives, the right ventricle and outflow tract, originate from prespecified early mesoderm just after gastrulation. A subset of the this area itself, initially called the anterior heart field, is marked by the presence of Fgf10 and Mef2c mRNA. This remarkable discovery has fundamentally altered the view of cardiac morphogenesis. Instead of continued growth of a defined population of differentiated cells, uncommitted cardioblasts from the second lineage are actively recruited into the heart, where they differentiate into cardiac cells. The discovery of the second lineage is also significant from the stance of a disease. To understand lesions involving the outflow tract and right ventricle, we must understand how the factors that regulate their formation from their precursors are coordinated and integrated with the rest of the heart. In DiGeorge syndrome, the defective gene operates primarily in the second lineage, affecting its differentiation and migration. Thus understanding of how the switch in lineage occurs from undifferentiated myoblasts from the second field to differentiated cardiac myocytes has many implications for embryology and disease.
The lineage of the second field has also been shown to be a multipotent precursor cell, which can give rise to all cardiovascular cell types, including myocytes, endothelial cells, and smooth muscle of the vasculature. These multipotent cells differentiate into the three different cell types presumably in response to local cues, such as growth factors, which instruct a particular gene program to be activated over another. In addition, these cardiovascular precursors restrict their potential as they further differentiate. This is a strategy similar to that used by hematopoietic precursors, which give rise to the different cell types that form blood.
The primary heart field was initially thought to contribute to the entire heart. Lineage analysis conclusively shows that it is the left ventricle and both atria that derive from the primary field. A subpopulation of cells defined very early in development by the expression of Tbx5 , prior to any sign of organogenesis, are already programmed to give rise to the left ventricle and atria, with the adjacent population giving rise to the right ventricle and outflow tract, as described above. This indicates that there is very early determination of future anatomically restricted precursors that are then patterned as the organ forms. Indeed, the junction between the first and second heart field derivatives results in a sharp boundary at the junction of the left and right ventricles, bisecting the interventricular septum. An even more refined patterning is evident in a sublineage that contributes solely to the left portion of the interventricular septum. One could imagine that this precise cellular arrangement would be critical for septum formation, and that its dysregulation might be at the root of ventricular septation defects.
Readers are encouraged to view interactive animated guides to heart morphogenesis at http://pie.med.utoronto.ca/HTBG/HTBG_content/assets/applications/index.html .
Genes That Regulate Formation of the Heart
Beginning with the discovery of the tinman mutant in Drosophila in 1989, several dozen genes have been identified that are critical for various aspects of formation of the heart, from its earliest inception, through major morphogenetic steps, and into postnatal regulation of cardiac function. Most genes encode transcriptional regulators, which turn on or off other genes, or signaling molecules that activate potent intracellular signaling cascades.
Transcription Factors
The fruit fly tinman mutation was identified in flies that did not form any heart at all, the mutation being named after the character in “The Wizard of Oz.” This mutation was in a gene belonging to a family of transcription factors called the homeodomain factors. Shortly after this discovery, vertebrate versions were identified, which were given the less colorful name of Nkx2-5 . It turns out that Nkx2-5 in vertebrates is not essential in itself for formation of the heart, but it does have important functions in early initiation of the cardiac genetic program, and in formation of the cardiac chambers. As discussed below, along with many other genes that were initially experimentally defined, NKX2-5 is one of the genes that has been identified as causative in inherited human congenitally malformed hearts. Considerable literature exists on the function of Nkx2-5 , and in many of its functions it interacts with other transcription factors that are important for the normal development of the heart. For example, a factor from another gene family, Gata4 , also plays important roles in heart differentiation, in chamber morphogenesis, and also has additional roles in bringing the two heart fields together. The last role was dramatically evident from its mutation in the mouse, which led to production of a bifid heart. And as with NKX2-5 , human genetics has pinpointed GATA4 as a gene that when mutated causes inherited congenital cardiac defects. The primary role of Nkx2-5 in the developing heart is to activate a set of target genes that will execute the correct cellular differentiation program of a variety of types of cells. For example, certain contractile protein genes rely on Nkx2-5 for their initial activation, and the proper development of the conduction system relies on its appropriate function.
Gata4 , as mentioned earlier, is also a key player in formation of the heart. Indeed, this gene has been shown to be important for such diverse aspects of formation as early differentiation, valvar formation, chamber maturation, and even postnatal function. In fact, these same roles have also been ascribed, to varying degrees, to Nkx2-5 . Genes that are active in the early heart usually have binding sites for Gata4 and Nkx2-5 in the regulatory regions, called “enhancers,” that confer cardiac-specific expression of genes. Gata4 and Nkx2-5 function together to act on these sequences of enhancers, and this interaction provides a degree of robustness and specificity to the system. Many other transcription factors have been defined as important for various aspects of cardiac formation. They often have multiple roles at various times during development, reflecting their potency and versatility.
Growth Factors
Growth factors of many different families are important for several aspects of heart development. In early development, the bone morphogenic factors and the Wnts , two types of developmentally important secreted factors, are key inducers of cardiac development, via their instructive cues that promote expansion of early cardiovascular precursors, and later also the induction of cardiac differentiation from these same precursors. The picture is a bit more complicated, as some Wnt signals also dampen cardiogenesis, by slowing the growth of precursors, presumably so that the timing of cardiac differentiation is kept.
Growth factors are also important in later stages of cardiac development, such as valvar formation and septation. Bone morphogenetic proteins are critical for the initiation of the earliest steps in valvar formation, and indeed the dosage of Bmp4 , for example, results in valvar malformations and deficient atrioventricular septation that are reminiscent of human disease. Later regulation of valvar morphogenesis relies on a complex interplay between myocardium and endocardium, which is regulated by calcineurin-dependent signaling and the repression of vascular endothelial growth factor in myocardium of the valve-forming region.
Micro-Rna Regulation of Cardiac Development
The transcriptional regulation of cardiac development, and its modulatory and instructive signaling pathways, are well studied, and their biology is becoming well understood. Less clear is the translational control of cardiac morphogenesis by small noncoding RNAs, such as microRNAs. MicroRNAs are genomically encoded 20–22 nucleotide RNAs that function by targeting mRNAs either for translational inhibition or for degradation, leading to an effective reduction in quantity of the protein product. Several hundred human microRNAs have been identified, and some of these have important roles in development that may be eminently relevant to congenitally malformed hearts.
The best characterized example is the microRNA-1 family, comprising miR-1-1 and miR-1-2 . These microRNAs are highly conserved from worms to humans, and are specifically expressed in the progenitor cells of developing cardiac and skeletal muscle as they differentiate. Both are highly expressed in the cells of the outflow tract derived from the second heart field. Interestingly, expression of these microRNAs is directly controlled by well-studied transcriptional regulatory networks that promote muscular differentiation. Consistent with a role in differentiation, overexpression of miR-1 in the developing mouse heart results in a decrease in expansion of ventricular myocytes, with fewer proliferating cardiomyocytes remaining in the cell cycle.
Defects caused by mutations in microRNA genes range from benign to severe. Disruption of the single fly orthologue of miR-1 had catastrophic consequences, resulting in uniform lethality at embryonic or larval stages, with a frequent defect in maintaining cardiac gene expression. Targeted deletion of miR-1-2 in the mouse resulted in ventricular septal defects, although with incomplete penetrance. In surviving adults, disruption of normal cardiac conduction and cell cycle control were also observed. As miRNAs can be highly redundant, deletion of all copies of redundant miRNAs must be accomplished to uncover their underlying function. Combined loss of miR-1-1 and miR-1-2 in the mouse indeed unveiled a profoundly important function of the miR-1 pair in broadly repressing a smooth muscle gene program, while promoting sarcomere formation in the developing heart. Many other miRNAs are enriched in specific cardiovascular cell types and play important roles in cardiogenesis, but in each case, they appear to be embedded in critical transcriptional networks, typically reinforcing the cellular actions of those networks (reviewed in Cordes and Srivastava ).
Hemodynamics and Formation of the Heart
As the heart forms, it soon begins to beat and pump blood. This occurs at the early stages of cardiac looping, well before chambers have formed and separations between segments of the heart are established. It would seem intuitive that the physical forces of a beating heart would affect its morphologic development, but until recently this has been but a concept. In fact, hemodynamic forces are indeed important, and shape not only the normal development of the heart but produce the secondary defects associated with major structural congenital cardiac malformations.
The initial identification of a role for hemodynamics was in zebrafish, a simple model of cardiac development in which a single atrium connects to a single ventricle. By altering the flow at the inflow or outflow of the zebrafish heart, and imaging the structure and function of the heart, it was determined that altering flow within the heart led to abnormal cardiac looping and defects in formation of the cardiac cushions, indicating that normal intracardiac flow is a key regulator of cardiac morphogenesis. This was confirmed using zebrafish and mouse models in which embryonic contractility is impaired or absent. More recent experiments, again in zebrafish, have clearly shown that intrinsic defects in contractility of cardiac myocytes, in combination with external hemodynamic forces, are essential for the normal development of the heart.
The most intriguing results on hemodynamics and cardiac development have a direct connection to congenital malformations. It had been observed in a mouse model of laterality defects that include anomalies of the outflow tracts that formation of the arteries in the branchial arches is randomized between the left and right sides. Normally, the arches form initially in bilateral fashion, but the left-sided brachiocephalic artery regresses, leaving only a right-sided artery. It was determined that, in this mouse model, the flow to the branchial arches was randomized to both left and right sides, unlike the usual situation where flow is preferentially directed to the right side. With a combination of surgical and morphologic manipulations, it was shown that the altered flow was the likely culprit for the abnormal presence of a left-sided brachiocephalic artery. Thus a genetic defect, in this case involving Pitx2 , leads not only to direct defects in regions of the heart where the mutated gene is expressed, but produces morphologic defects secondary to altered flow. This finding has important implications for the understanding of the origin of congenital cardiac malformations.
Single Gene Defects
In the last few decades, human genetic studies have identified several genes that are mutated in those with inherited cardiac disease. All of these are important regulators of cardiac morphogenesis, supporting the concept that the cardiac malformations are primarily a disease of abnormal embryonic organogenesis. These findings have also cemented the notion that most lesions have a genetic origin. In addition, these discoveries have clearly demonstrated that unrelated lesions can be caused by the same genetic defect, resulting in a need to reexamine epidemiologic studies from a different perspective. In all cases, the mutations are autosomal dominant, as they affect only one of two alleles of the gene. As yet, it is not fully known how dominant mutations in these genes result in profound aberrations in formation of the heart. However, most are likely related to decreased dosage, while some mutations lead to gain of function effects. It is of paramount importance to understand the molecular consequences of disease-causing mutations, and the mechanism underlying these events, in order to be able to design in rational fashion nonsurgical therapeutic interventions for those with congenitally malformed hearts. The future of our understanding of the etiology of the lesions therefore lies in understanding how mutations in these important regulators lead to altered morphogenesis. A few lessons from select mutations are described below.
Syndromic Mutations
Mutations of TBX5 in Holt-Oram Syndrome
Some genetic mutations have been identified in rare syndromes that include congenital cardiac malformations. For example, mutations in TBX5 cause defects in the context of Holt-Oram syndrome. These defects are predominantly interatrial communications, usually within the oval fossa, muscular ventricular septal defects, and abnormalities of the conduction system, but mutations can also result in more diverse abnormalities, such as hypoplastic left heart syndrome, totally anomalous pulmonary venous connection, and common atrioventricular junction ( Fig. 4.1 ). The syndrome also includes defects in limb formation, specifically of the radial ray, the thumb, first digit, and radius. Most defects of the limbs involve the thumbs, and range in severity from mild triphalangeal thumbs, up to complete phocomelia, with defects of intermediate severity such as missing thumbs, and hypoplastic radius.
Mutations have been found throughout the TBX5 gene, although they predominantly cluster in the so-called “T-box” domain, which confers DNA-binding capabilities on the protein. Early studies supported the concept of genotype-phenotype correlation, whereby mutations in one part of the gene were predominantly associated with families with severe cardiac defects, while mutations in a different part of the gene were found in families that had a predominance of severe limb, but not cardiac, defects. Subsequent findings, however, discounted this hypothesis. It is now believed that the difference in manifestation of disease is more likely a result of genetic background.
The observation by clinical geneticists that limb anomalies often coincide with cardiac defects have led some investigators to investigate TBX5 mutations for several related disorders. This often does not yield positive correlations. Recent work has strived to define the clinical criterion for the Holt-Oram syndrome in relation to the presence of TBX5 mutations. Indeed, there is excellent concordance, as long as a strict definition is maintained for the components of the syndrome.
Modelling of Holt-Oram syndrome in the mouse has revealed interesting and important features of the consequences of haploinsufficiency on downstream target gene regulation. In particular, it is clear that some genes in the heart are exquisitely sensitive to the dosage of Tbx5, explaining the occasional severity of defects in mice or humans with reduced dosage of this gene. Some genes, such as the gap junction protein-encoding gene Gja5 , also known as connexin 40, are targets that may explain some aspects of Holt-Oram syndrome, such as, in this case, the atrioventricular block. In other cases, investigation of the mouse model has led to the intriguing observation that independent of structural lesions one can identify clear defects in diastolic function due to reduced Tbx5 activity. These functional deficits are directly due to reduced function of the calcium pump, related to reduced activity of the Serca2a gene, which is directly regulated by Tbx5. These results, observed in the mouse, correlate to human patients with Holt-Oram syndrome, indicating that specific defects in cardiac function can accompany structural lesions while being unrelated to them except for their underlying genetic cause.
Mutations of SALL4 in Okihiro Syndrome
Mutations in the Spalt gene SALL4 have been found to underlie Okihiro syndrome, which, as with Holt-Oram syndrome, affects the limbs as well as the heart, although it also has a broad spectrum of ocular, renal, and other defects. The two syndromes are sometimes confused clinically due to very similar defects involving the heart and limbs, but molecular genetic studies have clearly outlined the distinction between the two. Using mouse models, it has been shown that Sall4 and Tbx5 genetically and physically interact to pattern the heart and limbs, thus explaining the reason for the similarities in syndromes caused by mutations in either genes.
Disruptions of the RAS Signaling Network in Noonan’s Syndrome
Autosomal dominant gain-of-function mutations in PTPN11 , encoding the protein tyrosine phosphatase SHP2, cause Noonan’s syndrome, characterized by pulmonary stenosis, hypertrophic cardiomyopathy, and occasional atrioventricular valvar defects. Most recently, hypomorphic mutations in SOS1 , an essential RAS guanine nucleotide-exchange factor (Ras-Gef), was shown to enhance RAS-ERK activation, and this can account for as high as one-fifth of the cases of Noonan’s syndrome not explained by PTPN11 mutations. Recent evidence implicates epidermal growth factor signaling as an important regulator of late valvar remodeling. Loss or attenuation of EGFR/ErbB1 signaling results in preferential hypercellularity of arterial but not atrioventricular valves. The hyperplastic arterial valvar phenotype is augmented when crossed to mice heterozygous for a null mutation in Ptpnl1 . Deletion of the EGF ligand, heparin-binding epidermal growth factor, results in increased size of the endocardial cushions and ridges, along with the size and proliferation of cells in both the arterial and atrioventricular valvar leaflets.
Nonsyndromic Mutations
Mutations of NKX2-5 Cause Septation and Other Defects
Mutations in the NKX2-5 gene cause defects similar to those caused by mutations of TBX5 , namely deficiencies of atrial and ventricular septation, problems with conduction, hypoplasia of the left heart, and other anomalies such as tetralogy of Fallot or Ebstein’s malformation. Mutations were originally identified in two families with very distinctive clinical features, namely progressive atrioventricular block in the setting of interatrial communications. Subsequent studies showed that this unusual combination was not the only type of lesion that could be caused by the mutations. Indeed, mutations in NKX2-5 were identified in several additional families that had some members presenting with an interatrial communication and atrioventricular block, but others with ventricular septal defects in isolation, Ebstein’s malformation, and tetralogy of Fallot ( Fig. 4.2 ). These results were extended to sporadic cases, which in addition to several cases of interatrial communication within the oval fossa and tetralogy of Fallot, included rare sporadic cases of common arterial trunk, double-outlet right ventricle, congenitally corrected transposition, interrupted aortic arch, hypoplastic left heart syndrome, and aortic coarctation. This was perhaps not surprising to developmental biologists, who would expect a broad set of defects based on the pan-cardiac expression of the Nkx2-5 gene in the mouse, but it showed with definitive clarity that a single defined mutation could result in a wide range of clinically and anatomically unrelated defects. This finding went a long way to explain the considerable difficulties in identifying familial inheritance of congenital cardiac malformations.
